Confidence in metabolite identification dictates the applicability of metabolomics to regulatory toxicology

Abstract

The strength of omics technologies in toxicology is their ability to identify pathways of toxicity. Although pathway discovery is not required for chemical risk assessment, it can generate molecular biomarkers that accelerate the use of molecular assays in regulatory toxicology. For metabolomics, this application is limited by the challenge of metabolite identification. Here, we construct a framework around the confidence in identification of metabolites and molecular pathways and map multiple toxicometabolomics studies to this framework. This reveals that few studies achieve the highest level of confidence defined by the Metabolomics Standards Initiative. However, we argue that the level of confidence required is dependent on the regulatory application. For some, such as chemical grouping, current practices can suffice, whereas other applications require a more rigorous approach to metabolite identification.